The rapid, error-free identification of microorganisms plays a prominent role in clinical microbiology in particular, but also in food analysis, monitoring and control of biotechnological processes, and monitoring of rivers and lakes. Microorganisms, which are also referred to as germs and microbes below, are generally microscopic organisms, which include bacteria, unicellular fungi (e.g. yeasts), microscopic algae and protozoa.
Identifying a microorganism means classifying it in the taxonomic hierarchical scheme: domain, kingdom, phylum, class, order, family, genus, species, and subspecies. The identification of bacteria, however, can additionally encompass varieties, such as serotypes or pathovars.
The term serotype or serovar (short for serovariety) is used to describe varieties within subspecies of bacteria which can be differentiated with serological tests. They differ in respect of antigens on the surface of the cells, and are identified in conventional microbiology with the aid of specific antibodies. The taxonomic hierarchy for serotypes is as follows: genus>species>subspecies (subsp.)>serotype, for example with the extended binomial species name Salmonella enterica subsp. enterica serotype Typhi, short form Salmonella Typhi. 
A pathovar (from the Greek pathos, meaning “suffering” or “disease”) is a bacterial strain or group of strains with the same properties that is differentiated from other strains within the species or subspecies on the basis of its pathogenicity. Pathovars are designated by means of a ternary or quaternary extension to the binomial species name. The bacterium Xanthomonas axonopodis, for example, which can cause citrus canker, has various pathovars with different host specializations: X. axonopodis pv. citri is one of them. The abbreviation “pv.” stands for “pathovar”. The virulent strains of human pathogens also have pathovars, but in this case they are designated by an addition before the name. For example, the intestinal bacterium Escherichia coli, which is mostly completely harmless, has the highly dangerous pathovars enterohaemorrhagic E. coli (EHEC), enteropathogenic E. coli (EPEC), enterotoxigenic E. coli (ETEC), enteroinvasive E. coli (EIEC), enteroaggregative E. coli (EAEC) and diffusely adherent E. coli (DAEC). The pathovars, in turn, can contain different serotypes. EHEC has many known serotypes, with around 60 percent of all identified EHEC serotypes being O157, O103 and O26. Particularly dangerous is the serosubtype O157/H7.
In a broader sense, the identification of microbes can also encompass varieties which differ in other medically relevant properties, in particular their resistance to antibiotics (especially beta-lactam antibiotics and glycopeptide antibiotics), but also their toxin formation (“toxivars”) or their receptiveness to the same or similar bacteriophages (“phagovars”). In general, the term “biovars” is used if a group of microbes of one species or subspecies have biological properties in common. One example of an antibiotic-resistant variety is MRSA: Methicillin-resistant Staphylococcus aureus. 
The term “strain” describes a population that was grown from a single organism and is kept in a (often national) repository for microbial strains. An internationally standardized strain designation is added to the nomenclature chain, comprising genus, species, subspecies and variety. The individual organisms of a strain are genetically identical; different strains vary slightly in their genetic make-up.
Two spectrometric methods have recently become widely used in microbiological laboratories for the identification of microbes. One of these is mass spectrometry (MS), and the other is infrared spectrometry (IR). It must be noted here that, strangely enough, no research group so far uses the two methods in parallel, or indeed in combination. On closer examination, this can be explained by the fact that the research fields of these groups are different in most cases, as are the aims of the identifications.
It is always favorable to use mass spectrometry when microbes of a completely unknown species have to be identified quickly and easily, down to the taxonomic level of the species, with no prior knowledge whatsoever. In general, the method works very well as long as the microbes can be cultured on or in nutrient media. It is preferable to produce colonies on gelatinous nutrient media in Petri dishes. The method is very robust: the age and nutrition of the colony have practically no effect on the mass spectrometric identification; nor are the quantities, preparation methods or storage periods of the samples on the mass spectrometric sample supports of any great importance. This means that the method of sample propagation and preparation does not require any strict standardization. Moreover, only very little sample material is needed, so tiny colonies are sufficient. The peaks of the mass spectra indicate microbe proteins that are exceedingly common or easy to ionize; 60 to 85 percent of these peaks belong to the 40 to 60 different proteins making up the ribosomes. Since each ribosome occurs several thousand times in each cell, and since the masses of the ribosomal proteins for the different microbe species are all characteristically different from each other, the uniform incidence of these proteins makes them ideal for an identification. After successful cultivation, the method identifies the taxonomic species of the microbes under investigation in a matter of minutes by using automated computer programs to compare their mass spectra with the reference mass spectra of an extensive spectral library, which may contain thousands of reference mass spectra of microbes across all taxonomic domains. The method has a very high identification certainty. In only a few cases is it not possible to differentiate between two microbe species with certainty. On the other hand, it is rarely possible to identify subspecies or even varieties, and, according to current knowledge, this will scarcely change on the basis of the methods currently in use, which are optimized for sensitivity, speed and identification accuracy. The principal reason for this is that the varieties do not differ in respect of their ribosomal proteins.
With infrared spectrometry, in contrast, it is possible to identify subspecies and varieties such as serotypes and pathovars, in some cases even individual strains within a species, if a suitable library of reference spectra for the subspecies and varieties of this species is available. The compiling of this library, however, usually requires that species-specific culture, preparation and measuring specifications are accurately adhered to, which is quite different to the situation with mass spectrometry.
The IR spectra are based on thousands of vibrations of the functional groups and the polar bonds in the biological material; these in turn originate from all the components of the microbial cells such as DNA, RNA, proteins, internal structures, membranes, and cell walls, right through to energy stores. There are no unequivocal assignments of molecules to individual characteristics in the spectra, albeit certain spectral ranges can be preferably assigned to certain molecular species: the fatty acid range from 3050 to 2800 cm−1 with vibrations of the CH2 and CH3 groups, the amide range from 1750 to 1500 cm−1 with peptide bonds, the polysaccharide range from 1200 to 900 cm−1. The range from 900 to 700 cm−1 is called the “fingerprint range”. It contains something from all the molecules and is very important for differentiating between the species.
The identifications depend on tiny differences in the IR spectra, which is why all method steps for identification via IR spectra are usually standardized, from the cultivation of the microbes with prescribed durations and temperatures on prescribed nutrient media through to the preparation and measurement of the samples. Oxygen content and moisture level above the nutrient medium must be controlled. Small deviations from the standardized method, such as a deviation in the culture period of more than half an hour or the culture temperature of more than one Kelvin, are enough to make the identification more difficult or falsify it.
In order to compile a library of reference IR spectra, microbes of a selected group are cultured and measured under standardized conditions. The spectra of this reference library are then subjected all together to a mathematical-statistical classification procedure. Several mathematical-statistical methods are used, such as ANN (artificial neural network analysis), PCA (principal component analysis), PLS-DA (partial least-square discriminant analysis), SVM (support vector machines), hierarchical cluster analyses or other classification techniques. After a learning and verification phase with the aid of the IR spectra from the library, the classification algorithms can then be applied to the infrared spectrum of the microbes of a sample, which were cultured in the same way. The algorithms provide the taxonomic classification, such as species, subspecies, serotype, pathovar and even strain. If, however, the microbe in the sample does not belong to the closely related group, the detailed IR analysis can provide totally incorrect results. As an example, the pathovar EHEC may be falsely indicated if the microbes are not E. coli, as assumed, but relatively harmless Citrobacter, for example.
These differences in handling and results mean that the fields of application for MS and IR are different. In clinical diagnostics, only mass spectrometry is used in practice, because unknown pathogens must be assumed across all taxonomic domains. The same applies to the monitoring of rivers and lakes and in all other areas where a fast identification of any type of microbes without any prior knowledge of the microbiological species is of paramount importance. The microbes here can belong to all the taxonomic domains of bacteria, archaea and eukaryotes, including unicellular algae or fungi, such as yeasts.
In contrast, if the aim is to detect sources of contamination and transmission routes of microbes in contaminated food or infected livestock, it is important to determine subspecies and varieties for a reliable identification of the infection sources. In this case, it can usually be assumed that the species of the microbes is known, at least after a single microbiological determination. Nowadays, therefore, IR spectrometry is used mainly where the species of the microbes is known, but it is important to accurately determine the subspecies and possibly the variety. If, for example, endemic salmonella poisoning occurs, and if it is suspected that the salmonella originate from an aquaculture of Thai shrimps, it is not sufficient to detect salmonella in the Thai shrimps. According to current taxonomy, there are two species of salmonella, one of which (Salmonella enterica) has five subspecies, but these have around 2600 different serovars as varieties between them. To come back to the shrimps, one then has to prove that the same salmonella variety is involved. The salmonella from a stool sample are therefore grown in a salmonella-specific culture broth, and IR spectrometry is used to examine the serotype of the selectively grown salmonella. This serotype must then be traced back to the Thai aquaculture.
Thus, infrared spectrometry is to be found mainly in food production, veterinary medicine and public health authorities, whereas clinical diagnostics is dominated by mass spectrometry.
Nowadays, the mass spectrometric identification method is also used around the world for medical diagnostics; in European and many other countries, methods and mass spectrometers from individual companies have already been clinically approved, as they meet the appropriate legal stipulations. Statutory approvals are being prepared in other countries. In the mass spectrometric method, the microbes are first cultured to form colonies. The nutrient medium for the culture is usually in an agar in a Petri dish, and by this method the cultivation of pure “isolates” in separate microbe colonies is achieved in hours or days, depending on the vigor of the microbes. It is not absolutely imperative to grow the microbes on agar, however. They can also be grown in liquids. If the colonies superimpose or mix, it is possible to obtain isolated colonies, again in the usual way, in a second cultivation. The tiny quantity of 104 to 106 microbes, hardly visible to the naked eye, which is transferred from a selected colony onto the mass spectrometric sample support by means of a small swab (preferably with a hygienically clean wooden toothpick), is then sprinkled with a strongly acidified solution of a conventional matrix substance (usually α-cyano-4-hydroxycinnamic acid, HCCA, or 2,5 di-hydroxybenzoic acid, DHB) for a subsequent ionization by matrix-assisted laser desorption (MALDI). The acid (usually formic acid or trifluoroacetic acid) attacks the cell walls, and the organic solvent (usually acetonitrile) of the matrix solution can penetrate into the microbial cells and cause their weakened cell walls to burst osmotically. The sample is then dried by evaporating the solvent, causing the dissolved matrix material to crystallize out. The soluble proteins of the microbes, and also other substances of the cell to a very small extent, are incorporated into the matrix crystals in the process.
The matrix crystals with the incorporated analyte molecules are bombarded with focused UV laser pulses in a mass spectrometer, thus generating ions of the analyte molecules in the vaporization plasmas. These ions can then be measured in the mass spectrometer, separated according to their mass. Simple time-of-flight mass spectrometers are commonly used for this purpose. The mass spectrum is the profile of the mass values of these analyte ions, which are very predominantly protein ions. The ions with the most useful information in terms of an identification have masses of between approximately 3,000 daltons and 15,000 daltons (1 dalton=1 atomic mass unit). In this method, the protein ions are very predominantly only singly charged (charge number z=1), which is why one can also simply talk about the mass m of the ions here, instead of always using the term “mass-to-charge ratio” mlz, as would otherwise actually be necessary in mass spectrometry.
Instead of simple time-of-flight mass spectrometers, it is also possible to use other types, such as time-of-flight mass spectrometers with orthogonal ion injection; and instead of the ionization by MALDI, it is certainly possible to use other types of ionization, such as electrospraying (ESI), although they provide more complicated mass spectra. A different ionization method for the generation of simple mass spectra is chemical ionization (CI), which can be used with laser-desorbed plasmas, for example.
The mass-separated profile of the soluble proteins, i.e. the mass spectrum, is very characteristic of the microbe species concerned because every species of microbe produces its own, genetically determined proteins, each having a characteristic mass. As has already been mentioned, around 60 to 85 percent of the proteins originate from the ribosomes. These are complexes of DNA and proteins which always have the same structure and which always contain between 40 and 60 different species-specific proteins in precisely the same number and composition. Each bacterial cell contains several thousand, and up to ten thousand, ribosomes; cells of eukaryotes contain several hundred thousand ribosomes. This means that not only the masses, but also the incidences of these soluble, more highly concentrated proteins are genetically predetermined; they do not depend on the nutritional conditions or the maturity of the colony, as do the lipoproteins, or the fatty acids which act as energy stores, for example. The protein profiles, especially those of the ribosomal proteins, are similarly characteristic of a microbe species as fingerprints are of an individual person. Reference libraries with reference mass spectra for thousands of microbes, which are legally approved for use in medical applications, are now available.
This mass spectrometric method of identification has proven to be extremely successful. The certainty of a correct identification is far greater than with the microbiological identification methods currently in use. In various studies it has been possible to prove that, with hundreds of different species of microbe, the identification certainty was far greater than 95 percent, and usually more than 98 percent. In cases of doubt, where there were deviations from current microbiological identification methods, genetic sequencing has confirmed that the mass spectrometric identification was correct in the vast majority of cases.
To identify the microbes, mass spectra from around 2,000 daltons up to high mass ranges of 20,000 daltons, for example, are measured, but it has been found that the mass signals in the lower mass range up to around 3,000 daltons can be evaluated less well because they can originate from substances whose presence is essentially random and variable, such as fatty acids, which are stored according to the nature of the nutrition. The best identification results are obtained by evaluating the mass signals in the mass range from around 3,000 to 15,000 daltons. The ultra-sensitive mass spectrometers now used for this purpose have only a low mass resolution, which means that the isotope groups, whose mass signals each differ by one dalton, cannot be resolved in this mass range. Only the envelopes of the isotope groups are measured.
This method of identifying microbes in principle requires a pure culture of microbes, a so-called “isolate”, in order to obtain a mass spectrum that is free of superimposed signals of other microbes. It has, however, been found that mass spectra of mixtures of two microbe species can also be evaluated, and that both microbe species can be identified. The identification certainty suffers only slightly. If more than two microbe species are involved in the mass spectrum, or if these two microbe species are present in very different concentrations, the identification probability and identification certainty decrease considerably.
Microbe identification by IR spectrometry is also based on pure cultures of microbes on suitable nutrient media. Here, however, age- and nutrition-related differences in the microbes must be avoided by maintaining standard conditions, since all components of the cells contribute to the IR spectra in all wavelength ranges in each case. The microbes, which are grown on standardized agar under standardized conditions, are suspended in pure water and deposited on an IR-transparent support plate. Care must be taken to ensure that the microbes are deposited in a uniform layer. The layer is dried and the absorption of the microbes on the support plate is measured in an infrared spectrometer. A Fourier Transform spectrometer (FT-IR), which has a high resolving power, is normally used. The spectra typically measured range from 4000 cm−1 to 500 cm−1. Several hundred spectra are measured and summed at acquisition rates of 20 spectra per second in order to improve the signal-to-noise ratio.
In a slightly modified embodiment, the IR spectra can also be measured in reflected light, in which case they are prepared on a metallically reflective support, made of aluminum, for example. It is also possible to use Raman spectroscopy, which has the advantage that the microbe spectra can also be measured in liquids.
There are other fields, besides food inspection and veterinary medicine, where there is a need to classify microbes in as much detail as possible according to subspecies and variety, but in these other fields the species of the microbe is usually completely unknown initially. In medical diagnostics, for example, knowledge about the pathogenicity, toxicity, virulence, and particularly the antibiotic resistance of the microbes is extremely important. These properties can certainly be very different for different subspecies or varieties of one microbe species.
The subspecies, serotypes, pathovars and further variations of the microbes are determined from their microbiological characteristics, for example from their attachment behavior (serotypes), their toxicity, their pathogenicity (pathovars), their virulence, and also from their resistance or non-resistance to the different antibiotics. There is (as yet) no detailed knowledge about which of these variations can be differentiated spectrometrically.
In view of the foregoing, there is a need to identify microbes from a wide range of taxonomic classes, where the classification should also extend in particular to manifestations below the taxonomic level of the species, i.e. subspecies, pathovars, toxivars, serotypes, and especially resistance to antibiotics.